Thread: BUG #15697: Restarting the server because of the error checkpointer

The following bug has been logged on the website:

Bug reference:      15697
Logged by:          Alex Newray
Email address:      aleksandr.newray@gmail.com
PostgreSQL version: 11.2
Operating system:   Ubuntu 16.04.6 LTS
Description:

I have a website using postgresql on my vps. I get an error restarting
postgresql server:

2019-03-15 05:58:53.305 UTC [1329] LOG:  shutting down
2019-03-15 05:58:53.313 UTC [1311] LOG:  checkpointer process (PID 1329)
exited with exit code 0
2019-03-15 05:58:53.313 UTC [1311] LOG:  terminating any other active server
processes
2019-03-15 05:58:53.314 UTC [6441] retouch@retouch_db WARNING:  terminating
connection because of crash of another server process
2019-03-15 05:58:53.314 UTC [6441] retouch@retouch_db DETAIL:  The
postmaster has commanded this server process to roll back the current
transaction and exit, because another
 server process exited abnormally and possibly corrupted shared memory.
2019-03-15 05:58:53.314 UTC [6441] retouch@retouch_db HINT:  In a moment you
should be able to reconnect to the database and repeat your command.
2019-03-15 05:58:53.316 UTC [6440] retouch@retouch_db WARNING:  terminating
connection because of crash of another server process
2019-03-15 05:58:53.316 UTC [6440] retouch@retouch_db DETAIL:  The
postmaster has commanded this server process to roll back the current
transaction and exit, because another
 server process exited abnormally and possibly corrupted shared memory.
2019-03-15 05:58:53.316 UTC [6440] retouch@retouch_db HINT:  In a moment you
should be able to reconnect to the database and repeat your command.
2019-03-15 05:58:53.318 UTC [1332] WARNING:  terminating connection because
of crash of another server process
2019-03-15 05:58:53.318 UTC [1332] DETAIL:  The postmaster has commanded
this server process to roll back the current transaction and exit, because
another server process exi
ted abnormally and possibly corrupted shared memory.
2019-03-15 05:58:53.318 UTC [1332] HINT:  In a moment you should be able to
reconnect to the database and repeat your command.
2019-03-15 05:58:53.320 UTC [1311] LOG:  all server processes terminated;
reinitializing
2019-03-15 05:58:53.341 UTC [13043] LOG:  database system was shut down at
2019-03-15 05:58:53 UTC
2019-03-15 05:58:53.348 UTC [1311] LOG:  database system is ready to accept
connections


I do load testing of the site and make a stream of requests directly to the
database and it works well. But for 2 days in a row at a certain time I get
an error restarting the server, and it does not depend on the load on the
server

The server has enough memory to start the server. The charts show that the
memory runs out, and in dmesg there is no record of postgres

PG Bug reporting form <noreply@postgresql.org> writes:
> I have a website using postgresql on my vps. I get an error restarting
> postgresql server:

> 2019-03-15 05:58:53.305 UTC [1329] LOG:  shutting down
> 2019-03-15 05:58:53.313 UTC [1311] LOG:  checkpointer process (PID 1329)
> exited with exit code 0
> 2019-03-15 05:58:53.313 UTC [1311] LOG:  terminating any other active server
> processes

Taking this at face value, it says that the checkpointer child process
exited before it was supposed to.  Given that it exited with exit(0),
that seems to mean that something sent it SIGUSR2, which is the
normal-shutdown signal that the postmaster would have sent it later on.
It's difficult to say anything beyond that on this much evidence,
but I suspect that there are moving parts you haven't told us about.
Are you using a nonstandard shutdown script?  Maybe there are poorly-
debugged extensions installed?

            regards, tom lane